Urgent.News

What's breaking now, across thousands of outlets.

AI

Lets talk about llms

Differently. (ChatGPT and LLMs—a friendly reminder) (Assume everything is possible—another friendly reminder.) LLMs are great. LLMs are easy and efficient. Yeah? But for how long? Let’s reject the premise that everything goes right. Let’s talk about it going wrong. What if it does go wrong? Seriously — what happens when we go wrong? Notice I said we . (One last friendly reminder.) How far can we…

LLMs, or large language models, are powerful tools that can generate human-like text with ease and efficiency. However, there are potential risks and unintended consequences to their usage. While it may be tempting to use an LLM to generate content for personal gain, such as creating a five-page website for a property, doing so can have negative impacts.

For example, the generated content could be scraped and used in future training data, potentially poisoning the model's perception of certain individuals, businesses, or properties. This could be particularly damaging for lesser-known people or properties with limited online presence.

Moreover, relying heavily on LLMs for information could lead to a false sense of credibility. If we become accustomed to using ChatGPT as our primary source of information, we may start to treat its responses as factual, even if they are not. This could have serious consequences for our relationships and decision-making processes.

In a worst-case scenario, if someone with significant resources and time were to use an LLM to manipulate its training data, it could potentially be used to spread misinformation or propaganda on a massive scale. This could be used to manipulate public opinion, distort reality, or even undermine societal norms.

While it may seem far-fetched, it is important to recognize that this is already happening to some extent. Political propaganda is not new, but with the advent of LLMs, the scale and speed at which misinformation can spread has never been greater. It is crucial that we remain vigilant and aware of the potential dangers of relying too heavily on LLMs for information, and take steps to ensure that we are not inadvertently contributing to the spread of falsehoods.

Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at dev.to →

More in AI

Superpowers for Coding Agents: Turn Vague Requests Into Tested Changes

The fastest way for an AI coding agent to create expensive work is to start coding too soon. A request arrives, the agent infers the missing requirements, selects an architecture, edits several files…

  • Superpowers methodology introduces gates to prevent costly coding mistakes.
  • Agent recognizes need to create something, then understands user intent.
  • Testing and subagent-driven development ensure final product aligns with user intent.

Beyond Words: Building an AI Mental Health Monitor with HuBERT and Psycho-Acoustics

We often focus on what someone says, but in the realm of clinical psychology, how they say it is often more revealing. Subtle changes in speech—a slight tremor (jitter), a slowing tempo, or a…

  • Mental health monitor uses HuBERT model for acoustic analysis
  • Preprocesses audio input and extracts F0, speech rate, jitter
  • Deploys model via ONNX Runtime for real-time monitoring

More from Sunday 23 August →